I am using the OpticalFlowPyrLk to track some point in an image. I know that the input point list is a set of floating point values. I normal use floating point values with no decimal part as the input (e.g. 125.00). If I use an input value with a decimal part (e.g. 125.639), can I expect to get a correct displacement based on that decimal part, or is the value truncated to an integer part before the calculations are done?
For more complete information about compiler optimizations, see our Optimization Notice.